• ANN: Lea 3.4.0 released

    From pie.denis@skynet.be@21:1/5 to All on Sat Nov 14 18:30:49 2020
    Lea 3.4.0 is now released!
    http://pypi.org/project/lea/3.4.0

    What is Lea?
    ------------
    Lea is a Python module aiming at working with discrete probability distributions in an intuitive way.

    It allows you modeling a broad range of random phenomena: gambling, weather, finance, etc. More generally, Lea may be used for any finite set of discrete values having known probability: numbers, booleans, date/times, symbols,.
    Each probability distribution is modeled as a plain object, which can be
    named, displayed, queried or processed to produce new probability distributions.

    Lea also provides advanced functions and Probabilistic Programming (PP) features; these include conditional probabilities, joint probability distributions, Bayesian networks, Markov chains and symbolic computation.

    All probability calculations in Lea are performed by a new exact algorithm,
    the Statues algorithm, which is based on variable binding and recursive generators. For problems intractable through exact methods, Lea provides on-demand approximate algorithms, namely MC rejection sampling and
    likelihood weighting.

    Beside the above-cited functions, Lea provides some machine learning
    functions, including Maximum-Likelihood and Expectation-Maximization algorithms.

    Lea can be used for AI, education (probability theory & PP), generation of random samples, etc.
    LGPL - Python 2.6+ / Python 3 supported

    For a 5 minutes tour. check out the poster presented at PROBPROG2020 conference:
    http://probprog.cc/assets/posters/fri/69.pdf


    What's new in Lea 3.4.0?
    ------------------------
    Lea 3.4.0 includes two important improvements over 3.3.x:
    1) Introduction of "evidence contexts", allowing to factorize conditions
    when calculating conditional probabilities

    http://bitbucket.org/piedenis/lea/wiki/Lea3_Tutorial_3#markdown-header-evide nce-contexts-evidence-add_evidence-methods
    2) Optimize calculation for several queries, which were intractable in an
    exact way with previous Lea versions

    http://bitbucket.org/piedenis/lea/issues/60/scaling-exact-inference-optimiza tion

    This version contains also a couple of improvements on
    usability/consistency.

    To learn more...
    ----------------
    Lea 3 on PyPI -> http://pypi.org/project/lea
    Lea project page -> http://bitbucket.org/piedenis/lea
    Documentation -> http://bitbucket.org/piedenis/lea/wiki/Home
    Statues algorithm -> http://link.springer.com/chapter/10.1007/978-3-030-52246-9_10

    With the hope that Lea can make this universe less erratic,

    Pierre Denis

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)